Artificial intelligence procedures are disclosed for localizing faults in corrupted messages in 5G and 6G, and for correcting those faults based on measured parameters such as backgrounds and message signals according to pulse-amplitude modulation. An AI model with multiple adjustable variables may be “trained” using a large number of message events, including faulted messages, to determine which message elements are likely faulted, based on input parameters such as modulation quality, SNR, and other signal properties. The receiving entity can then attempt a grid search to correct the faulted message elements, or request a retransmission. For field use by base stations and user devices, an algorithm may be developed based on the AI model, and configured to predict which message elements are likely faulted. By detecting and correcting message faults, networks may increase reliability and reduce latency while avoiding most retransmission costs and delays, according to some embodiments.
Legal claims defining the scope of protection, as filed with the USPTO.
2. The method of claim 1, wherein the message complies with 5G or 6G technology.
4. The method of claim 1, wherein the background property comprises a noise level or an interference level, or a combination thereof.
5. The method of claim 1, wherein the predicting comprises calculating each of the internal functions according to the input parameters, and then calculating the output result according to the internal functions.
6. The method of claim 1, wherein the output result comprises a list of predicted probabilities, each probability in the list indicating a probability that one of the message elements is faulted, respectively.
7. The method of claim 1, further comprising providing, as an input to the artificial intelligence array, a parameter of a demodulation reference, the demodulation reference being associated with the message.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
February 18, 2022
August 9, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.